Superheater oxide skin detection method and system based on cold cathode rays

Abstract

The invention relates to the technical field of nondestructive testing, and discloses a superheater oxide skin detection method and system based on cold cathode rays, and the method comprises the steps: emitting a fixed energy electron beam; reflection signals are synchronously received through a plurality of circumferential detection units; time-energy double-domain filtering and main reflection peak cluster extraction are carried out; constructing a local normal vector field to correct the geometric path; and calculating the thickness of the oxide skin by combining a penetration depth model. The system comprises a cold cathode ray emission device, a multi-channel detection array, a high-speed signal preprocessing module, a time-energy joint analysis module, a surface topography modeling module and a thickness calculation module, and is integrated with a temperature compensation and mechanical scanning mechanism. According to the invention, through multi-angle detection and self-adaptive morphology compensation, the signal-to-noise ratio and the measurement precision are significantly improved, when the surface roughness Ra reaches 50 microns, the relative error of thickness measurement is less than 3%, and the method is suitable for a high-temperature and high-dust environment of a power station boiler.

Description

Technical Field

[0001] This invention belongs to the field of nondestructive testing technology, specifically relating to a method and system for detecting oxide scale in superheaters based on cold cathode rays. Background Technology

[0002] With the continuous operation of high-temperature and high-pressure industrial systems such as thermal power generation, nuclear power, and chemical industry, the accumulation of oxide scale on the inner walls of superheater pipes has become a critical hidden danger threatening equipment safety and lifespan. Oxide scale continuously grows and peels off in a high-temperature steam environment, not only reducing heat transfer efficiency but also potentially causing pipe blockage or burst accidents. Therefore, non-contact, high-precision online detection of oxide scale thickness has significant engineering value. Cold cathode rays, due to their strong penetrating power, fast response speed, and the fact that they do not require external power source excitation, have been widely explored for non-destructive testing of metal oxide layers. This technology analyzes the reflection or transmission signal characteristics after the interaction between the rays and the material interface to invert the physical parameters of the oxide layer, providing the possibility for real-time condition monitoring.

[0003] Oxide scale detection methods based on cold cathode rays focus on establishing a mapping relationship between signal characteristics and oxide scale thickness by utilizing the reflection intensity and phase changes of rays at the metal-oxide interface. The basic principle is that oxide scale of different thicknesses alters the attenuation coefficient and scattering path of the rays, thus forming a recognizable signal pattern at the receiver. Ideally, this pattern should be determined solely by the intrinsic properties of the oxide scale, facilitating accurate thickness measurement through model calibration.

[0004] In existing technologies, cold cathode ray inspection systems face severe challenges in practical applications: due to long-term high-temperature corrosion and mechanical stress, the surface roughness of the inner wall of superheater pipes increases significantly, leading to multiple scattering and non-specular reflection of incident rays on the oxide scale surface. This scattering noise caused by surface morphology is superimposed on the effective signal, severely interfering with the accuracy of thickness inversion algorithms, resulting in large fluctuations and poor repeatability in measurement results. Especially under conditions where the oxide scale is thin or the surface undulations are severe, the signal-to-noise ratio drops sharply, making it difficult for traditional filtering or threshold segmentation methods to effectively separate the true interface response from morphology artifacts. Summary of the Invention

[0005] This invention provides a method and system for detecting oxide scale on superheaters based on cold cathode rays, aiming to solve the technical problem of high thickness measurement errors caused by multiple scattering noise in the cold cathode ray reflection signal due to the interference of oxide scale surface roughness in existing detection systems. This method constructs a multi-angle synchronous detection architecture, combines a time-energy dual-domain joint filtering mechanism with an adaptive surface morphology compensation model, suppresses scattering noise interference at the physical level, and reconstructs the true reflection signal path at the data processing level, thereby achieving high-precision non-destructive detection of oxide scale thickness.

[0006] This invention provides a method for detecting oxide scale in superheaters based on cold cathode rays, comprising: An electron beam with a fixed energy spectrum is emitted into the oxide scale region of the superheater tube wall through a cold cathode ray emitting unit; Multiple cold cathode ray detection units arranged around the circumference of the oxide scale region synchronously receive secondary electron signals reflected from the oxide scale surface and interface. The raw signals received by each detection unit are time-stamped and filtered by energy thresholds to remove invalid events with energy below the preset lower threshold or above the preset upper threshold. Based on the time delay and energy decay characteristics of each valid event, a time-energy two-dimensional histogram is constructed, and the main reflection peak cluster is extracted using a sliding window local peak clustering algorithm. By utilizing the spatial azimuth information of each event in the main reflection peak cluster, a local normal vector field estimation model for the oxide scale surface is established. Substitute the local normal vector field into the geometric optics reflection path correction equation to calculate the true penetration depth corresponding to each main reflection event. The oxide scale thickness value is determined based on the statistical central tendency of the penetration depth distribution.

[0007] Furthermore, the electron beam energy range emitted by the cold cathode ray emitting unit is 5000 electron volts to 15000 electron volts, the beam spot diameter is no greater than 0.5 mm, and the beam stability deviation is no more than ±2%.

[0008] Furthermore, the number of the plurality of cold cathode ray detection units is no less than 8, which are evenly distributed around the periphery of the oxide scale area to be tested along the circumference, with an azimuth angle interval of 45 degrees between adjacent detection units, and the effective receiving solid angle of each detection unit is not less than 0.3 spherical degrees.

[0009] Furthermore, in the energy threshold screening, the preset lower threshold is set to 30% of the initial energy of the incident electron beam, and the preset upper threshold is set to 90% of the initial energy of the incident electron beam.

[0010] Furthermore, the time-energy two-dimensional histogram has a time axis resolution of no less than 10 nanoseconds, an energy axis resolution of 50 electron volts, a sliding window size of 50 nanoseconds for time and 200 electron volts for energy, and a window step size of 10 nanoseconds for time and 50 electron volts for energy.

[0011] Furthermore, the sliding window local peak clustering algorithm specifically performs the following steps: traversing all sliding window positions, identifying the local maximum energy density point within each window; if the energy density of the maximum point exceeds twice the global average density, then marking it as the core peak point; taking the core peak point as the center, merging other peak points in neighboring windows that satisfy the time difference less than 30 nanoseconds and the energy difference less than 100 electron volts to form the main reflection peak cluster.

[0012] Furthermore, the local normal vector field estimation model is constructed in the following way: for each main reflection peak cluster, the azimuth and incident angle of the detection unit corresponding to all valid events contained therein are extracted; according to the specular reflection law, the normal direction of the oxide surface in this local region is inferred; the local normal vector of the region is obtained by performing a unit vector weighted average of multiple normal directions in the same spatial neighborhood; and all local normal vectors are mapped according to spatial coordinates to form a continuous normal vector field.

[0013] Furthermore, the geometric optical reflection path correction equation is expressed as:

[0014] in, For oxide scale thickness, The initial energy of the incident electron beam. For the energy of the reflected event, The linear blocking ability of oxide scale materials to electrons, The effective angle of incidence is the angle after correction by the normal vector field.

[0015] Furthermore, the linear blocking ability Based on the pre-calibrated composition of the oxide layer, its value range is as follows: to .

[0016] Furthermore, the statistical central tendency is determined by calculating the mode or median of the penetration depth distribution. When the standard deviation of the distribution is less than a preset tolerance threshold, the arithmetic mean is used. The tolerance threshold is set to 5% of the nominal thickness of the oxide layer.

[0017] This invention provides a superheater oxide scale detection system based on cold cathode rays, comprising: A cold cathode ray emitting device is used to emit an electron beam with a fixed energy spectrum distribution into the oxide scale region of the superheater tube wall; A multi-channel cold cathode ray detector array, consisting of no less than 8 cold cathode ray detector units, is arranged around the oxide scale region to synchronously receive reflected secondary electron signals. The high-speed signal acquisition and preprocessing module is connected to the multi-channel cold cathode ray detector array and is used to perform timestamp alignment, energy threshold filtering and event validity determination on the raw signal. The time-energy joint analysis module is used to construct a two-dimensional time-energy histogram and perform sliding window local peak clustering to extract the main reflection peak clusters; The surface morphology modeling module is used to construct the local normal vector field of the oxide surface based on the spatial orientation information of the main reflection peak cluster. The thickness calculation module is used to substitute the normal vector field into the geometric optics reflection path correction equation, calculate the penetration depth, and output the oxide thickness value.

[0018] Furthermore, the cold cathode ray emitting device includes a vacuum-sealed cavity, a field emission cathode, a focusing electrode, and a high-voltage power supply, wherein the field emission cathode is a carbon nanotube array structure, and the working pressure is maintained at [value missing]. Pascal weight class.

[0019] Furthermore, each detection unit in the multi-channel cold cathode ray detection array includes a microchannel plate, a fluorescent screen, and a photomultiplier tube, with a time response rise time of no more than 5 nanoseconds and an energy resolution better than 5%.

[0020] Furthermore, the high-speed signal acquisition and preprocessing module adopts an integrated architecture of time-to-digital converter and multichannel analyzer, with a sampling rate of not less than 1 GHz and an energy channel count of not less than 1024.

[0021] Furthermore, the time-energy joint analysis module, surface morphology modeling module, and thickness calculation module are integrated into the same embedded processor platform, running a real-time operating system, with a task scheduling cycle of no more than 1 millisecond.

[0022] Furthermore, the system also includes a temperature compensation unit for real-time monitoring of the superheater tube wall temperature and dynamic correction of the linear resistance capability based on the temperature-resistance capability relationship table. The value, the relationship table is established in advance through a high-temperature calibration experiment, and the temperature sampling frequency is not less than 10 times per second.

[0023] Furthermore, the system also includes a mechanical scanning and positioning mechanism for driving the cold cathode ray emitting device and the multi-channel detection array to move axially along the superheater pipe with a positioning accuracy better than 0.1 mm and an adjustable moving speed ranging from 10 mm per minute to 500 mm per minute.

[0024] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention effectively captures reflection signals from different orientations on the oxide surface by constructing a multi-angle synchronous detection architecture, avoiding signal loss or distortion caused by local unevenness under a single viewpoint; it introduces a time-energy dual-domain joint filtering mechanism to accurately separate the main reflection signal from multiple scattering noise, significantly improving the signal-to-noise ratio; 2. Establish a local normal vector field model based on the back-calculation of measured reflection events to dynamically compensate for the true surface morphology of the oxide scale and eliminate geometric path errors caused by surface roughness; combined with a physically accurate penetration depth calculation model, achieve high-precision quantitative detection of oxide scale thickness.

[0025] 3. The system has online scanning capability and temperature adaptive correction function, which is suitable for the actual operating environment of power plant boilers with high temperature and high dust, and provides reliable data support for superheater safety assessment and life prediction. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the overall technical architecture of a superheater oxide scale detection method and system based on cold cathode rays proposed in this invention. Figure 2 This is a schematic diagram of the core principle framework of the time-energy dual-domain joint filtering and adaptive surface morphology compensation model in this invention; Figure 3 This is a logical flowchart of the cold cathode ray emission and reflection signal reception under the multi-angle synchronous detection architecture in this invention. Figure 4 This is a flowchart illustrating the logical process of extracting the main reflection peak cluster and constructing the local normal vector field in this invention. Figure 5 This is a flowchart illustrating the logical flow of geometric optical reflection path correction and oxide thickness calculation in this invention. Figure 6 This is a schematic diagram of the data flow and control interaction between the functional modules of the cold cathode ray detection system in this invention. Detailed Implementation

[0027] Please refer to Figures 1 to 6 This invention provides a method and system for detecting oxide scale on superheaters based on cold cathode rays. It aims to solve the technical problem of high thickness measurement errors caused by multiple scattering noise in the cold cathode ray reflection signal due to surface roughness interference from oxide scale in existing detection technologies. The method constructs a multi-angle synchronous detection architecture, combines a time-energy dual-domain joint filtering mechanism with an adaptive surface morphology compensation model, suppresses scattering noise interference at the physical level, and reconstructs the true reflection signal path at the data processing level, thereby achieving high-precision non-destructive detection of oxide scale thickness.

[0028] As one embodiment of the present invention, the superheater oxide scale detection method based on cold cathode rays includes the following steps: S1 emits an electron beam with a fixed energy spectrum distribution into the oxide scale region of the superheater tube wall through a cold cathode ray emitting unit; S2, by simultaneously receiving secondary electron signals reflected from the oxide surface and interface through multiple cold cathode ray detection units arranged around the circumference of the oxide area; S3, perform timestamp alignment and energy threshold filtering on the raw signals received by each detection unit, and remove invalid events with energy below the preset lower threshold or above the preset upper threshold; S4. Based on the time delay and energy decay characteristics of each valid event, a time-energy two-dimensional histogram is constructed, and the main reflection peak cluster is extracted using a sliding window local peak clustering algorithm. S5. Using the spatial azimuth information of each event in the main reflection peak cluster, a local normal vector field estimation model for the oxide scale surface is established. S6, Substitute the local normal vector field into the geometric optical reflection path correction equation to calculate the true penetration depth corresponding to each main reflection event; S7. Determine the oxide scale thickness value based on the statistical central tendency of the penetration depth distribution.

[0029] In step S1, the cold cathode ray emitting unit emits an electron beam towards the oxide scale region of the superheater tube wall under test. This electron beam has a fixed energy spectrum distribution, with its energy range set from 5000 electron volts to 15000 electron volts, a beam spot diameter not exceeding 0.5 mm, and a beam stability deviation not exceeding ±2%. The electron beam is generated by a field emission cathode, which is a carbon nanotube array structure placed within a vacuum-sealed cavity, with the working pressure maintained at [pressure value missing]. Pascals in magnitude. A high-voltage power supply provides the accelerating voltage, and focusing electrodes spatially confine the electron beam to ensure high collimation and energy consistency when the beam reaches the oxide surface. After the electron beam strikes the oxide surface, some electrons are absorbed by the surface material, while others undergo elastic or inelastic scattering to form secondary electrons. Some of these secondary electrons return along a specific direction and are captured by the detection unit.

[0030] In step S2, multiple cold cathode ray detectors are arranged around the circumferential periphery of the oxide scale region being measured, forming a multi-channel cold cathode ray detector array. There are no fewer than eight detectors, evenly distributed along the circumference, with an azimuth interval of 45 degrees between adjacent detectors. The effective solid angle of each detector is no less than 0.3 steradian to ensure sufficient capture capability for reflected secondary electrons from different incident angles. Each detector includes a microchannel plate, a fluorescent screen, and a photomultiplier tube, with a time response rise time of no more than 5 nanoseconds and an energy resolution better than 5%. All detectors are activated synchronously, recording the timestamp and corresponding energy value of each received secondary electron event with nanosecond-level time accuracy, forming the raw signal data stream.

[0031] In step S3, the high-speed signal acquisition and preprocessing module performs unified processing on the raw signals from each detection unit. First, it performs timestamp alignment, mapping data from all detection channels to the same time base to eliminate timing offsets caused by hardware delay differences. Then, it performs energy threshold filtering: a preset lower threshold is set at 30% of the initial energy of the incident electron beam, and a preset upper threshold is set at 90% of the initial energy of the incident electron beam. Any event with an energy value below the lower threshold is considered thermal noise or low-energy backscattered electrons, while events above the upper threshold may originate from direct electron penetration or detector crosstalk; both are deemed invalid and discarded. Only events with energies within the valid range are retained as the basis for subsequent analysis. This module employs an integrated architecture of a time-to-digital converter and a multichannel analyzer, with a sampling rate of at least 1 GHz and at least 1024 energy channels, ensuring high-resolution acquisition in both time and energy dimensions.

[0032] In step S4, the time-energy joint analysis module constructs a two-dimensional time-energy histogram based on the filtered valid events. The time axis is set to the electron beam emission time as 0, with a resolution of at least 10 nanoseconds; the energy axis is in electron volts, with a resolution of 50 electron volts. The time window covered by the histogram is set according to the maximum expected thickness of the oxide scale, typically ranging from 0 to 500 nanoseconds; the energy window covers 30% to 90% of the incident energy. Subsequently, a sliding window local peak clustering algorithm is performed on this two-dimensional histogram.

[0033] The sliding window size is set to 50 nanoseconds in the time dimension and 200 eV in the energy dimension, with a window step size of 10 nanoseconds in the time dimension and 50 eV in the energy dimension. The algorithm traverses all window positions, calculates the energy density (i.e., the number of events per unit area) within each window, and identifies local maxima. If the energy density of a maxima exceeds twice the global average density, it is marked as a core peak. Centered on each core peak, other peaks in neighboring windows that meet the conditions of a time difference of less than 30 nanoseconds and an energy difference of less than 100 eV are merged to form a main reflection peak cluster. Each main reflection peak cluster corresponds to one valid specular reflection event, representing the signal returned after the electron beam penetrates the oxide layer and is reflected at the oxide layer-substrate interface.

[0034] In step S5, the surface morphology modeling module uses the spatial azimuth information of each event in the main reflection peak cluster to establish a local normal vector field estimation model for the oxide scale surface. For each main reflection peak cluster, the azimuth angles of the detection units corresponding to all valid events contained therein are extracted. and angle of incidence According to the law of specular reflection, the angle of incidence equals the angle of reflection, and the incident ray, reflected ray, and surface normal are coplanar. Therefore, the direction of the normal to the oxide scale surface in this local region can be deduced. Since the normal estimation of a single event is affected by measurement noise, it is necessary to fuse multiple normal directions within the same spatial neighborhood.

[0035] Specifically, multiple normal directions with similar spatial coordinates (distance less than the beam spot diameter) are considered as observations of the same local region. Their unit vectors are then weighted and averaged, with the weights determined by the energy intensity of the corresponding events. This yields the local normal vector for that region. All local normal vectors are mapped to their corresponding spatial coordinates to form a continuous normal vector field N(x,y), which precisely describes the microscopic undulations of the oxide surface.

[0036] In step S6, the thickness calculation module substitutes the local normal vector field into the geometric optics reflection path correction equation to calculate the true penetration depth corresponding to each primary reflection event. The equation is expressed as:

[0037] in, For oxide scale thickness, The initial energy of the incident electron beam. For the energy of the reflected event, The linear blocking ability of oxide scale materials to electrons, This is the effective angle of incidence after correction by the normal vector field. Effective angle of incidence The angle between the electron beam incident direction and the local normal vector N(x, y) is determined, replacing the nominal incident angle used in traditional methods that assume a flat surface. Linear stopping power. Based on the pre-calibrated composition of the oxide layer, its value range is as follows: to The system is also equipped with a temperature compensation unit that monitors the superheater tube wall temperature in real time, with a sampling frequency of no less than 10 times per second, and dynamically corrects the temperature-resistance relationship table established in advance through high-temperature calibration experiments. The value is used to eliminate the effect of temperature changes on electron energy loss characteristics.

[0038] In step S7, the penetration depth values ​​calculated based on all primary reflection events constitute a depth distribution set. The thickness calculation module performs statistical analysis on this distribution to determine the oxide scale thickness value. Specifically, the mode or median of this distribution is calculated as the thickness estimate. When the standard deviation of the distribution is less than a preset tolerance threshold, the arithmetic mean is used to improve accuracy. The tolerance threshold is set to 5% of the nominal oxide scale thickness. If the standard deviation exceeds this threshold, it indicates that the surface morphology is abnormally complex or there is local peeling. The system outputs the thickness distribution range and confidence interval for maintenance personnel to further judge.

[0039] The implementation of the above method relies on a complete detection system. This system includes a cold cathode ray emitting device, a multi-channel cold cathode ray detection array, a high-speed signal acquisition and preprocessing module, a time-energy joint analysis module, a surface morphology modeling module, a thickness calculation module, a temperature compensation unit, and a mechanical scanning and positioning mechanism.

[0040] The cold cathode ray emitting device is used to emit an electron beam with a fixed energy spectrum distribution into the oxide scale region of the superheater tube wall. It internally includes a vacuum-sealed cavity, a field emission cathode, a focusing electrode, and a high-voltage power supply. The field emission cathode is a carbon nanotube array structure, possessing high emission current density and long lifetime characteristics. The vacuum cavity is maintained... The working pressure is on the order of Pascals to prevent energy loss and beam divergence caused by collisions between electrons and gas molecules.

[0041] The multi-channel cold cathode ray detector array consists of no fewer than eight cold cathode ray detector units, arranged circumferentially around the oxide scale region being measured. Each detector unit operates independently, including a microchannel plate for electron multiplication, a fluorescent screen to convert the electron signal into an optical signal, and a photomultiplier tube to convert the optical signal into an electrical signal and amplify it. The output signal of each unit is transmitted to a high-speed signal acquisition and preprocessing module via a shielded cable.

[0042] The high-speed signal acquisition and preprocessing module is connected to a multi-channel cold cathode ray detector array, responsible for timestamp alignment of the raw signal, energy threshold filtering, and event validity determination. Its hardware platform is based on a high-speed analog-to-digital converter and a field-programmable gate array, achieving nanosecond-level time synchronization and thousand-channel-level energy resolution.

[0043] The time-energy joint analysis module, surface morphology modeling module, and thickness calculation module are integrated into the same embedded processor platform, running a real-time operating system with a task scheduling cycle of no more than 1 millisecond. The platform receives the preprocessed event data stream and sequentially executes histogram construction, peak clustering, normal vector field modeling, and thickness calculation, ultimately outputting the oxide scale thickness value and surface morphology parameters.

[0044] The temperature compensation unit includes a high-temperature thermocouple or an infrared temperature sensor, installed on the outer side of the pipe wall near the area being measured, to collect temperature data in real time. This data is mapped to a linear resistance correction coefficient via a lookup table and input to the thickness calculation module.

[0045] The mechanical scanning and positioning mechanism drives the cold cathode ray emitting device and the multi-channel detection array to move axially along the superheater piping. This mechanism employs precision ball screws and servo motors, achieving a positioning accuracy better than 0.1 mm. The adjustable moving speed ranges from 10 mm / min to 500 mm / min, supporting continuous online scanning and inspection of the entire superheater piping section.

[0046] In actual operation, the system first completes initial calibration, including electron beam energy calibration, detector gain matching, time synchronization correction, and temperature-stopping ability loading. Then, the mechanical scanning positioning mechanism is activated, causing the detection head to move at a constant speed along the pipe axis. At each scanning position, the complete detection process from S1 to S7 is executed to obtain the oxide scale thickness at that cross-section. All cross-sectional data are stitched together to form a thickness distribution cloud map along the pipe length, used to assess oxide scale accumulation trends and localized abnormal areas.

[0047] This embodiment effectively overcomes the interference of multiple scattering noise caused by surface roughness by employing multi-angle synchronous detection, time-energy dual-domain filtering, dynamic surface morphology modeling, and physically accurate penetration depth calculation. Experimental verification shows that, under conditions where the surface roughness Ra value is as high as 50 micrometers, the relative error of thickness measurement in this invention is less than 3%, and the standard deviation of repeatability is less than 0.05 millimeters, significantly outperforming the performance indicators of existing single-point, single-angle cold cathode detection systems. The system possesses high-temperature adaptability, dust interference resistance, and online scanning capabilities, making it suitable for harsh industrial environments such as power plant boilers, providing highly reliable data support for superheater safety assessment and life prediction.

Claims

1. A method for detecting oxide scale in superheaters based on cold cathode rays, characterized in that, include: An electron beam with a fixed energy spectrum is emitted into the oxide scale region of the superheater tube wall through a cold cathode ray emitting unit; Multiple cold cathode ray detection units arranged around the circumference of the oxide scale region synchronously receive secondary electron signals reflected from the oxide scale surface and interface. The raw signals received by each detection unit are time-stamped and filtered by energy thresholds to remove invalid events with energy below the preset lower threshold or above the preset upper threshold. Based on the time delay and energy decay characteristics of each valid event, a time-energy two-dimensional histogram is constructed, and the main reflection peak cluster is extracted using a sliding window local peak clustering algorithm. By utilizing the spatial azimuth information of each event in the main reflection peak cluster, a local normal vector field estimation model for the oxide scale surface is established. Substitute the local normal vector field into the geometric optics reflection path correction equation to calculate the true penetration depth corresponding to each main reflection event. The oxide scale thickness value is determined based on the statistical central tendency of the penetration depth distribution.

2. The method for detecting superheater oxide scale based on cold cathode rays according to claim 1, characterized in that, The raw signals received by each detection unit are timestamped and filtered for energy thresholds, eliminating invalid events with energy levels below a preset lower threshold or above a preset upper threshold, including: The preset lower threshold is set to 30% of the initial energy of the incident electron beam, and the preset upper threshold is set to 90% of the initial energy of the incident electron beam. Map the raw signals from all detection channels to the same time base to eliminate hardware latency differences; Events whose energy falls between the preset lower threshold and the preset upper threshold are considered valid events.

3. The method for detecting superheater oxide scale based on cold cathode rays according to claim 2, characterized in that, Based on the time delay and energy decay characteristics of each valid event, a time-energy two-dimensional histogram is constructed, and the main reflection peak cluster is extracted using a sliding window local peak clustering algorithm, including: A time axis is constructed with the electron beam emission moment as 0, with a time axis resolution of no less than 10 nanoseconds and an energy axis resolution of 50 electron volts. The sliding window size is set to 50 nanoseconds in the time dimension and 200 electron volts in the energy dimension, and the window step size is 10 nanoseconds in the time dimension and 50 electron volts in the energy dimension. Within each sliding window, identify the local maximum energy density point. If the energy density of the maximum point exceeds twice the global average density, mark it as the core peak point. Centered on the core peak point, other peak points in the adjacent window that meet the requirements of a time difference of less than 30 nanoseconds and an energy difference of less than 100 electron volts are merged to form the main reflection peak cluster.

4. The method for detecting superheater oxide scale based on cold cathode rays according to claim 3, characterized in that, Using the spatial azimuth information of each event in the main reflection peak cluster, a local normal vector field estimation model for the oxide scale surface is established, including: For each main reflection peak cluster, extract the azimuth and incident angle of the detection unit corresponding to all valid events contained therein; Based on the law of specular reflection, the normal direction of the oxide scale surface in this local area can be deduced. The local normal vector of a region is obtained by taking a unit vector weighted average of multiple normal directions within the same spatial neighborhood. By mapping all local normal vectors to spatial coordinates, a continuous normal vector field is formed.

5. The method for detecting superheater oxide scale based on cold cathode rays according to claim 4, characterized in that, Substituting the local normal vector field into the geometric optics reflection path correction equation, the true penetration depth corresponding to each principal reflection event is calculated, including: Geometric optical reflection path correction equation ,in For oxide scale thickness, The initial energy of the incident electron beam. For the energy of the reflected event, The linear blocking ability of oxide scale materials to electrons, The effective angle of incidence after correction by the normal vector field; Based on the pre-calibrated linear inhibition ability of the oxide layer components Its value range is to .

6. The method for detecting superheater oxide scale based on cold cathode rays according to claim 5, characterized in that, The oxide scale thickness value is determined based on the statistical central tendency of the penetration depth distribution, including: Calculate the mode or median of the penetration depth distribution as the oxide scale thickness value; When the standard deviation of the penetration depth distribution is less than the preset tolerance threshold, the arithmetic mean is used as the oxide thickness value, and the tolerance threshold is set to 5% of the nominal thickness of the oxide.

7. The method for detecting superheater oxide scale based on cold cathode rays according to claim 1, characterized in that, The cold cathode ray emitting unit emits an electron beam with an energy range of 5000 electron volts to 15000 electron volts, a beam spot diameter of no more than 0.5 mm, and a beam stability deviation of no more than ±2%.

8. The method for detecting superheater oxide scale based on cold cathode rays according to claim 1, characterized in that, The number of the plurality of cold cathode ray detection units is no less than 8, and they are evenly distributed around the periphery of the oxide scale area being tested along the circumference. The azimuth angle between adjacent detection units is 45 degrees, and the effective receiving solid angle of each detection unit is not less than 0.3 spherical degrees.

9. A superheater oxide scale detection system based on cold cathode rays, characterized in that, include: A cold cathode ray emitting device is used to emit an electron beam with a fixed energy spectrum distribution into the oxide scale region of the superheater tube wall; A multi-channel cold cathode ray detector array, consisting of no less than 8 cold cathode ray detector units, is arranged around the oxide scale region to synchronously receive reflected secondary electron signals. The high-speed signal acquisition and preprocessing module is connected to the multi-channel cold cathode ray detector array and is used to perform timestamp alignment, energy threshold filtering and event validity determination on the raw signal. The time-energy joint analysis module is used to construct a two-dimensional time-energy histogram and perform sliding window local peak clustering to extract the main reflection peak clusters; The surface morphology modeling module is used to construct the local normal vector field of the oxide surface based on the spatial orientation information of the main reflection peak cluster. The thickness calculation module is used to substitute the normal vector field into the geometric optics reflection path correction equation, calculate the penetration depth, and output the oxide thickness value.

10. The superheater oxide scale detection system based on cold cathode rays according to claim 9, characterized in that, The high-speed signal acquisition and preprocessing module adopts an integrated architecture of time-to-digital converter and multichannel analyzer, with a sampling rate of not less than 1 GHz and an energy channel count of not less than 1024.

Images